DUIANet: A double layer U-Net image hiding method based on improved Inception module and attention mechanism

Xintao Duan, Guoming Wu, Chun Li, Zhuang Li, Chuan Qin

Abstract: Highlights•We propose a novel double-layer U-Net structure that can fully extract and fuse image features, generate high-quality image, and still have good performance when hiding multiple images.•Our method improves the anti-steganalysis ability of steganographic image. By cascading convolution and attention mechanism, secret image can be hidden in more suitable locations when hiding image, improving the security of steganographic image.•Our method has strong generalization ability. The improved Inception module expands the network width, extract features from multiple scales, thereby fusing feature information from different dimensions.